Morpho vs Aave V3: Rates, Risk Model and Liquidations Compared

Aave V3 is a single shared liquidity pool where governance sets, and can later change by vote, every asset's risk parameters, while Morpho is a set of isolated, immutable, permissionless markets fixed at creation by five parameters, with curated vaults layered on top that spread a depositor across several of those markets under caps and timelocks. That single difference explains everything else below: how rates move, who you are trusting, what a liquidator can do in one transaction, and what you have to watch.

Architecture: one pool versus many markets

On Aave V3 you supply into a shared pool of per-asset reserves, each with its own LTV, liquidation threshold and interest rate configuration. Everything you supply in that market backs everything you borrow in it: the health factor is an account-wide number, computed with a weighted average liquidation threshold across all of your collateral.

Those parameters are not constants. Aave's documentation is explicit: "The parameters for each reserve are not fixed; they vary between markets and can change over time as Aave Governance monitors market conditions and adjusts settings accordingly." Votes cover things like raising borrow caps or adjusting LTVs, and the same token can carry different settings in different deployments.

Morpho goes the other way. A market is defined by five parameters, chosen once at creation:

• Collateral asset
• Loan asset
• LLTV, the liquidation loan-to-value
• Oracle
• Interest rate model (IRM)

Those markets are immutable ("Once created, rules never change"), isolated ("Each market operates independently") and permissionless ("New market doesn't require governance vote to be created"). Governance keeps two whitelists and nothing else: the approved LLTV values (0%, 38.5%, 62.5%, 77.0%, 86.0%, 91.5%, 94.5%, 96.5%, 98.0%) and the approved interest rate models, currently only the AdaptiveCurveIRM.

That is the point of isolated markets: risk is scoped to one collateral and loan pair, so a broken oracle or a collapsing collateral asset cannot reach the market next to it. On Aave the same event is absorbed inside the shared pool, where every supplier of that reserve sits on the same side of it.

It also changes what a depositor touches. On Aave you supply to a reserve and you are done. On Morpho you pick a market yourself, or you deposit into a curated vault, an ERC-4626 vault holding positions across several markets for you. That vault is what makes Morpho feel like Aave to a passive lender, with one difference the interface understates: the risk decisions are delegated to a named curator instead of to protocol governance.

Rates, and why the same asset prices differently

Aave V3 uses a two-slope kinked curve: "the model of interest rate is based on two slopes, one before the OPTIMAL_USAGE_RATIO point of usage and another from that point to 100%." The optimal usage ratio, both slopes, the base rate and the reserve factor are per-asset parameters set by governance, so an Aave curve stays exactly where it is until a vote moves it.

Morpho's AdaptiveCurveIRM is "engineered to maintain the ratio of borrowed assets over supplied assets, commonly called utilization, close to a target of 90%", combining a curve that responds immediately when utilization moves away from target with an adaptive mechanism that shifts the target rate itself when it stays away. The documented speeds:

• At 45% utilization, the rate progressively decreases until it is divided by 2 after 10 days
• At 95% utilization, it progressively increases until it doubles after 10 days
• At 100% utilization, it doubles after 5 days, the maximum speed

The curve is steep too. The docs work through a market at 4% with 90% utilization: the same market prices at 16% at 100% utilization and 1% at 0%. That gives three mechanical reasons a Morpho lending rate and an Aave one can differ for the same asset at the same moment:

• Different target utilization. Morpho aims for 90%. Aave's optimal usage ratio is a governance-set per-asset parameter and is typically lower, so a Morpho market can sit closer to fully utilized, passing more of what borrowers pay through to suppliers.
• The Morpho curve moves on its own. Aave's curve is fixed until a vote. Morpho's target rate drifts with sustained utilization, so the same utilization can produce a different rate a week later.
• Isolation concentrates demand. An Aave reserve blends all borrow demand for an asset at one rate. A Morpho market is one pair with its own utilization, so a busy pair prices independently of a quiet one holding the same loan asset.

Risk: governance on one side, curators and oracles on the other

On Aave, the dominant non-market risk is governance risk. The liquidation threshold inside your health factor is a parameter, and parameters get revoted. The upside is that one professional risk process covers every user of the pool. The downside is that the margin you sized when you opened the position is not guaranteed to be the one you are liquidated on.

Morpho inverts that. A market's rule never changes, so risk moves to who created it and what it points at:

• Oracle risk. Every market names its own oracle at creation, and Morpho puts it plainly: "No oracle is immune to price manipulation, which can lead to liquidations or even bad debt." A faulty one damages that market and every vault allocated to it.
• Curator risk. If you deposit into a vault rather than a market, the curator controls the strategy through adapter configuration and fee settings, and a compromised owner account can replace the curator.

Both share the rest: bad debt risk, when collateral falls below the borrowed amount before liquidators act; liquidity risk, since "a lack of liquidity can prevent suppliers from withdrawing their assets for a certain period of time"; and smart contract risk.

The red and yellow warnings in the Morpho app

Because market creation is permissionless, anyone can spin up a market with any collateral and any oracle. Morpho handles that in the interface with a three-tier warning system, triggered when there is an unrecognized parameter or a potential abnormality with a market:

• Red: any collateral asset, loan asset, deposit asset, oracle or vault curator not found on a recognized list. You must opt in before interacting.
• Yellow: a soft warning that flags a potential abnormality but does not require you to opt in.
• Blacklisted: markets created with tokens that violate trademarks, have proven to be scams, or are excluded based on legal considerations. Hidden from the interface entirely.

The warnings come from the Morpho risk API rather than being hard-coded in the frontend, so the set is dynamic. Red is a gate, not a block, and blacklisting is the only tier that removes a market from view.

Vault reallocation: what a curator can and cannot do with your deposit

A curated vault does not park your money in one market. It routes it. "The supply queue determines the order in which markets receive new deposits", each taking liquidity in sequence until the deposit is placed or the caps are reached, and "the withdraw queue determines the order in which markets are tapped to fulfill user withdrawals." The documented best practice is a supply queue of [Idle] and a withdraw queue of [Idle, Market 1, Market 2, ...], so deposits land idle first and get placed deliberately afterwards.

Placement is the allocator's job: it "uses allocate and deallocate to move assets between the vault's idle pool and the various enabled adapters", within curator-defined caps. What the curator sets:

• Which protocols, markets and asset types are permissible at all, and which adapters are enabled
• Absolute caps (a fixed amount) and relative caps (a percentage of vault assets) on risk identifiers such as a collateral type, an oracle or a protocol. The documented example: max total exposure to stETH as collateral, 15M
• Who the allocators are, and the fees, with a performance fee capped at 50% and a management fee capped at 5%

And what it cannot do:

• It cannot move liquidity day to day. That is the allocator, which cannot introduce new, unapproved risks, only shuffle between enabled adapters
• It cannot act instantly on most risk-increasing changes: "nearly all significant actions are subject to timelock, providing depositors an exit window." Cutting a cap, which reduces risk, is instant
• It can be overruled by the sentinel, which deallocates assets, decreases caps and revokes pending timelocked actions, and can only reduce risk

Timelocks are per function in Vault V2, configurable from 0 to 3 weeks, with documented examples of 7 days for high-risk actions and 1 day for routine ones. In V1 a single global timelock covers actions such as raising a market's supply cap, and must sit between 1 day and 2 weeks.

For a depositor that adds up to something concrete. Your exposure changes market by market without you signing anything, inside caps you can read in advance, and most risk-increasing changes come with a window to leave, which is only useful if you notice it. If withdrawals exceed the idle pool plus the liquidity adapter, the vault becomes illiquid until allocation changes. And forceDeallocate is permissionless: you can redeem vault receipt tokens directly for the underlying positions, with a penalty of up to 2%.

Liquidations compared

Both liquidate on the same principle, a debt that grew past what the collateral supports. They differ on the two parameters a borrower feels: how much debt can be closed at once, and how much the liquidator takes.

The close factor is the first genuine split. Aave caps a single liquidation at 50% of your debt in the normal case, leaving a partially repaired position behind. Morpho has no cap, so a liquidation can close the entire position in one transaction.

The incentive is the second. Aave's liquidation bonus is a governance parameter set per asset, so it can move. Morpho's Liquidation Incentive Factor is a pure function of the market LLTV, known at creation and fixed forever: about a 5% bonus in an 86% LLTV market, never more than 15%, and all of it to the liquidator with no protocol cut.

The blast radius differs too. On Aave, one falling collateral asset drags the whole account's health factor down. On Morpho, a borrower with three markets has three independent liquidation points and can lose one without the other two moving. The liquidation calculator works for both shapes, and health factor across lending protocols covers how the number itself is defined protocol by protocol.

Which one to use, honestly

Aave V3 fits if you want one account-wide number to track, cross-collateral efficiency, and a risk process you do not run yourself. You are trusting governance to keep parameters sane, and accepting that they can change under an open position.

Morpho fits if you want a rule that cannot be revoted, exposure contained to one pair, or, as a lender, a vault whose mandate and caps you can actually read. The cost is homework: a market is only as good as its oracle and its LLTV, and a vault is only as good as its curator, its caps and its timelocks. More to check, but checkable.

Neither is safer in the abstract. Aave concentrates decisions into one thing to watch. Morpho distributes them into several, each smaller. Plenty of people run both, borrowing wherever the borrowing rate is cheaper that week.

Monitoring both

Running both means watching two different things. On Aave it is one account-wide health factor that can move because the market moved or because a vote changed a parameter. On Morpho it is a per-market number whose rule never changes, sitting inside a vault whose allocation a curator can reshape under timelock.

Otomato detects your Aave and Morpho positions automatically from a wallet address, read-only, no wallet connect and no signatures, then alerts you on every event linked to them, on Telegram or in the mobile app: health factor moves, rate changes, security incidents, depegs on the assets you hold. If Aave is your main exposure, the Aave health factor alert is the one to set up first.

Clément Hecquet

CEO & Co-founder · Otomato

Co-founder and CEO of Otomato. Engineer turned product leader, building the portfolio-aware alerts layer for on-chain users.